Files
ragflow/internal/parser/parser/pptx_parser.go
Jack 9b0719fa94 fix: Go ingestion migration batch 5 (Parser 1.1/1.7/2.11, Chunker 1.7/1.8/2.6/2.7, Tokenizer 6x fixes) (#17419)
## Summary

Continuation of the Python→Go ingestion pipeline migration (File →
Parser → Chunker → Extractor → Tokenizer). Fixes cover Parser, Chunker,
and Tokenizer gaps identified. Fix page number (0-indexed and 1-index
mixed before fix; use 1-indexed after fix) and chunk order issues.

### Parser
- **Slides TCADP (1.7):** `pptx_tcadp.go` + TCADP branch in
`pptx_parser.go`/`ppt_parser.go` — PowerPoint files now support
`parse_method="tcadp"` via the TCADP cloud service, matching the
spreadsheet-family TCADP pattern. PPT containers pass `"PPT"` as
fileType (not hardcoded `"PPTX"`).
- **Audio default output_format (2.11):** `defaultSetups()` audio
default changed from `"text"` to `"json"`, aligning with Python
`parser.py:232` and `AllowedOutputFormat["audio"]={"json"}`.
- **PDF VLM enhancement (1.1):** `maybeDispatchPDFVisionEnhancement` in
`pdf_vision_dispatch.go` enriches image/table items with IMAGE2TEXT
model descriptions after PDF parsing, mirroring Python
`enhance_media_sections_with_vision`. Semaphore fix: acquire before
goroutine start to prevent unbounded goroutine creation.
- **json family (2.3):** reclassified as Keep Go — `json_parser.go` is a
functional enhancement, not a parity gap.
- **page number:** changed from "mixed use of 1-indexed & 0-indexed" to
"1-indexed"

### Chunker
- **BULLET_PATTERN fallback (1.7):** 4th-level fallback in
`resolveTitleLevels` (`title.go`) detects bullet/numbered-list patterns
(Chinese legal, numbering, English) when outline + regex levels produce
only bodyLevel. Guarded by `allBodyLevel` to never override existing
structure.
- **Tag/One chunker fields (1.8):** `tag.go` sets `TopInt` from source
row index; `one.go` preserves `Positions`/`PDFPositions` from source
items. TSV multi-line RowNum fix: tracks `contentStart` for correct row
attribution.
- **Overlapped_percent normalization (2.6):**
`NormalizeOverlappedPercent` in `schema/chunker.go` mirrors Python
`common/float_utils.py:50-58` — accepts `[0,1)` fraction or `[0,90]`
percent, normalizes to canonical `[0,90]`.
- **Paragraph splitting (2.7):** aligned to Python flow `naive_merge` —
`CRLF` normalization, `splitKeepingDelimiter` preserves sentence
delimiters, single-section merge with token-budget-governed chunking.
- **chunk order:** sort by reading order

### Tokenizer
- **Phantom chunk filtering (Omission 2):** `isPhantomChunk` + filter
loop in `chunksFromTokenizerUpstream` skips zero-value ChunkDocs.
- **Batch size env var (Omission 3):** `embeddingBatchSize()` reads
`TOKENIZER_EMBEDDING_BATCH_SIZE`, defaults to 16.
- **Summary empty check (Diff 5):** `TrimSpace(s) != ""` → `s != ""`,
matching Python truthy check.
- **chunk_order_int all paths (Diff 8):** set unconditionally before
full_text/embedding branching.
- **Timeout default (Diff 10):** `600s` → `60s`, matching Python
`@timeout(60)`.
- **Small maxTokens truncation (Diff 14):** `truncateForEmbedding`
returns `""` when `maxTokens <= 10`, matching Python.

### Code review fixes
- Semaphore acquire moved before goroutine in `pdf_vision_dispatch.go`
(concurrency control)
- Context propagation in `pptx_tcadp.go` (cancellation support)
- Test resolver leak fix in `media_dispatch_test.go` (defer restore)
- Migration history comments removed per AGENTS.md

## Test plan
```
bash build.sh --test ./internal/parser/parser/... ./internal/ingestion/component/...
```

## Notes
- Migration diff tracking: `docs/migration_python_go_diff.md`
- Remaining gaps: Extractor component only (21 items)
2026-07-28 11:12:52 +08:00

145 lines
4.3 KiB
Go

//go:build cgo
//
// Copyright 2026 The InfiniFlow Authors. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
//
package parser
import (
"context"
"fmt"
"strings"
officeOxide "github.com/yfedoseev/office_oxide/go"
)
// PPTXParser parses both .pptx (OOXML) and .ppt (OLE binary)
// files via the office_oxide backend. The format field controls
// the container format passed to OpenFromBytes — "pptx" for
// ZIP-based OOXML presentations and "ppt" for the legacy binary
// OLE format.
type PPTXParser struct {
format string
// TCADP cloud-parsing configuration
ParseMethod string
TCADPAPIServer string
TCADPAPIKey string
TCADPTableResultType string
TCADPMarkdownImageResponseType string
OutputFormat string
}
func NewPPTXParser() *PPTXParser {
return &PPTXParser{format: "pptx"}
}
func (p *PPTXParser) String() string {
return "PPTXParser"
}
// ConfigureFromSetup reads the slides-family setup map. Mirrors the
// XLSXParser ConfigureFromSetup pattern
func (p *PPTXParser) ConfigureFromSetup(setup map[string]any) {
if p == nil || setup == nil {
return
}
if v, ok := setup["parse_method"].(string); ok {
p.ParseMethod = v
}
if v, ok := setup["tcadp_apiserver"].(string); ok {
p.TCADPAPIServer = v
}
if v, ok := setup["tcadp_api_key"].(string); ok {
p.TCADPAPIKey = v
}
if v, ok := setup["table_result_type"].(string); ok {
p.TCADPTableResultType = v
}
if v, ok := setup["markdown_image_response_type"].(string); ok {
p.TCADPMarkdownImageResponseType = v
}
if v, ok := setup["output_format"].(string); ok {
p.OutputFormat = v
}
if p.OutputFormat == "" {
p.OutputFormat = "json"
}
}
// ParseWithResult emits one JSON item per slide with the slide's
// plain text. Mirrors the python parser.py:slides branch which
// forces output_format="json" for the slide family.
func (p *PPTXParser) ParseWithResult(ctx context.Context, filename string, data []byte) ParseResult {
// p == nil guard: the struct is embedded by value in PPTParser and
// always created via NewPPTXParser or the "ppt"-format constructor in
// PPTParser, so this branch is unreachable from normal call paths.
// Kept as defensive guard — a nil dereference here would obscure the
// root cause behind a nil-pointer panic.
if p == nil {
return ParseResult{Err: fmt.Errorf("PPTXParser is nil")}
}
method := strings.ToLower(strings.TrimSpace(p.ParseMethod))
switch method {
case "tcadp":
return parsePresentationWithTCADP(ctx,
filename, data, strings.ToUpper(p.format),
p.TCADPAPIServer, p.TCADPAPIKey,
p.TCADPTableResultType, p.TCADPMarkdownImageResponseType,
p.OutputFormat,
)
case "", "deepdoc":
// Continue with the local office_oxide parser.
default:
// PDF-specific methods like "paddleocr" / "mineru" are
// meaningless for PPTX; treat as default path.
}
doc, err := officeOxide.OpenFromBytes(data, p.format)
if err != nil {
return ParseResult{Err: fmt.Errorf("presentation open: %w", err)}
}
defer doc.Close()
text, err := doc.PlainText()
if err != nil {
return ParseResult{Err: fmt.Errorf("pptx plain-text: %w", err)}
}
// Split on form-feed (the python TxtParser convention used by
// RAGFlow's slide parser) — each block becomes a JSON item.
var items []map[string]any
for i, raw := range strings.Split(text, "\f") {
trimmed := strings.TrimSpace(raw)
if trimmed == "" {
continue
}
items = append(items, map[string]any{
"text": trimmed,
"doc_type_kwd": "text",
"slide_number": i + 1,
})
}
if items == nil {
items = []map[string]any{{"text": strings.TrimSpace(text), "doc_type_kwd": "text"}}
}
return ParseResult{
OutputFormat: "json",
File: map[string]any{"name": filename, "format": p.format},
JSON: items,
}
}